Web technologies created immense practical usage value right from the start. The fact that cryptocurrencies will do the same is very, very debatable. For now, the only valuable use case I know of is for illegal activities (ransomware, tax evasion, buying illegal stuff). Which is a sizeable market, but not one which is going to see mass adoption.
It's not all that different from the dot-com boom; I (mostly) missed out on the bubble from 1995-2000 since I was in high school at the time, but kept studying computers & web technologies in college and caught the mass adoption wave afterwards. I had a number of classmates and coworkers (at internships), slightly older than me, who rode the dot-com bubble right into the ground, losing both their jobs and their life savings. Many of them left tech altogether and became economists or traveled abroad or bounced around between temp work. They basically swore off investing in tech stocks, but one of the folks I'm thinking of thought she'd make a pretty penny by buying 3 houses and investing in hedge funds in 2007.
It's human nature to believe that bad things that happened to other people aren't going to happen to you, and as a species we're remarkably bad at learning from other people's experience. The set of folks actually burned by ICOs is a relatively small subset of "all people" though - IIRC at the peak of the bubble there were only about 10M Bitcoin investors in the U.S. and only a small percentage of them actually invested in ICOs. That's like 3% of the population, still very much in early-adopter territory.